Search Images Maps Play YouTube News Gmail Drive More »
Sign in
Screen reader users: click this link for accessible mode. Accessible mode has the same essential features but works better with your reader.

Patents

  1. Advanced Patent Search
Publication numberUS7038617 B1
Publication typeGrant
Application numberUS 11/138,109
Publication dateMay 2, 2006
Filing dateMay 26, 2005
Priority dateMay 26, 2005
Fee statusLapsed
Publication number11138109, 138109, US 7038617 B1, US 7038617B1, US-B1-7038617, US7038617 B1, US7038617B1
InventorsJohn G. Ferrante
Original AssigneeLockheed Martin Corporation
Export CitationBiBTeX, EndNote, RefMan
External Links: USPTO, USPTO Assignment, Espacenet
Sensor doppler/wavelet processing
US 7038617 B1
Abstract
A method according to an aspect of the invention is for filtering coherent multipulse radar, sonar or lidar returns. The method comprises the steps of separating the returns into mutually orthogonal I & Q components, performing wavelet processing on the I and Q components to generate an I & Q wavelet filter output. The I & Q wavelet filter outputs are data compressed to remove clutter portions. Data compression may be removal of large-amplitude or earliest wavelet responses. An inverse wavelet filtering function is performed on the compressed wavelet filter outputs, to thereby generate edited coherent I & Q components. Doppler filtering, is performed on the edited coherent I and Q components to thereby generate frequency-domain target data in which the minimum discernible target velocity is improved.
Images(2)
Previous page
Next page
Claims(4)
1. A method for filtering coherent multipulse radar return signals, said method comprising the steps of:
separating said return signals into mutually orthogonal I and Q components;
performing wavelet processing on said I component to generate an I wavelet filter output;
performing wavelet processing on said Q component to generate a Q wavelet filter output;
data compressing said I and Q wavelet filter outputs to remove those portions which are most likely to contain clutter to thereby generate data compressed wavelet filter outputs;
performing an inverse wavelet filtering function on the compressed wavelet filter outputs, to thereby generate edited coherent I and Q components;
Doppler filtering said edited coherent I and Q components to thereby generate frequency-domain target data in which the minimum discernible target velocity is improved.
2. A method according to claim 1, wherein said step of data compressing said I and Q wavelet filter outputs comprises the step of removing those portions of said I and Q wavelet filter outputs which have the largest amplitudes.
3. A method according to claim 1, wherein said step of data compressing said I and Q wavelet filter outputs comprises the step of removing those I and Q wavelet filter outputs which occur earliest in the stream of wavelet responses.
4. A method according to claim 1, further comprising the step of Doppler filtering said mutually orthogonal I and Q components into frequency bins having frequencies above the frequencies associated with said minimum discernible target velocity.
Description
FIELD OF THE INVENTION

This invention relates to sensors such as radar, sonar, and lidar, using a combination of wavelet and Doppler processing to enhance resolution of relatively low-speed targets in the presence of clutter.

BACKGROUND OF THE INVENTION

In radar (or sonar or lidar) processing to identify targets, the radar signals which are transmitted are ideally reflected by the desired target, and returned to the radar system for identification and ranging. One of the problems to which most radar systems are subject is that of clutter. Clutter is caused by undesired targets, such as trees in a land-based system or waves in a maritime system, which undesired targets are relatively near the radar system, or otherwise located so as to generate an undesired radar return signal. Radar system return signals are often said to follow an “inverse fourth-power” law, in that the strength of the return signal becomes smaller in proportion to the fourth power of the range of the target. A concomitant is that nearby clutter will produce target returns which are very much greater than the returns from small or inconspicuous targets at substantial range. Clutter can occur at any range, however, and tends to obscure desired targets at that range.

Schemes have been developed for reducing the effects of clutter. Moving-target indication (MTI) is a system which tends to suppress or cancel those return signals which exhibit the same return signal phase from pulse to pulse, which generally applies to nonmoving targets. Doppler filtering is a scheme in which the returned radar signal is applied to an array of frequency bins, which separate the signals in accordance with how fast the target is moving. The clutter signals tend to be grouped in the lowest-frequency bins, and can be processed independently from those returns showing higher frequencies indicative of greater speed.

Unfortunately, slow-moving targets, such as helicopters, which are near the radar system, may have return signals which are near the locations in which clutter occurs, and may have target velocities which are low or in the same range as that of clutter. Doppler filtering is not fully effective in extracting target information under such conditions.

Improved resolution of slow-moving targets is desired in the presence of clutter.

SUMMARY OF THE INVENTION

A method according to an aspect of the invention is for filtering coherent multipulse radar return signals. The method comprises the steps of separating the return signals into mutually orthogonal I and Q components, performing wavelet processing on the I component to generate an I wavelet filter output, and performing wavelet processing on the Q component to generate a Q wavelet filter output. The I and Q wavelet filter-outputs are data compressed to remove those portions which are most likely to contain clutter, to thereby generate data compressed wavelet filter outputs. An inverse wavelet filtering function is performed on the compressed wavelet filter outputs, to thereby generate edited coherent I and Q components. Doppler filtering, which may be in the form of the discrete Fourier transform or the fast Fourier transform (DFT or FFT), is performed on the edited coherent I and Q components to thereby generate frequency-domain target data in which the minimum discernible target velocity is improved.

In a particular mode of the method of this aspect of the invention, the step of data compressing the I and Q wavelet filter outputs comprises the step of removing those portions of the I and Q wavelet filter outputs which have the largest amplitudes.

In another particular mode, the step of data compressing the I and Q wavelet filter outputs comprises the step of removing those I and Q wavelet filter outputs which occur earliest in the stream of wavelet responses corresponding to the lowest scale or lowest frequency region in the Doppler domain.

BRIEF DESCRIPTION OF THE DRAWING

FIG. 1 is a simplified block diagram of a portion of a radar sensor according to an aspect of the invention.

DESCRIPTION OF THE INVENTION

FIG. 1 is a simplified block diagram of a portion of a radar receiver according to an aspect of the invention. In FIG. 1, radar return signals resulting from target reflections due to coherent transmitted pulses appear on an input signals path 12, and are applied to an intermediate-frequency (IF) receiver 14. Receiver 14 also receives local oscillator signals by way of a path 16, and produces analog inphase (I) and quadrature (Q) baseband signals, which are applied by way of individual path portions 18I and 18Q of a path 18 to an analog-to-digital converter (ADC) 20. ADC 20 converts the analog I and Q signals to digital form, and the resulting digital I and Q signals appear on paths 22I and 22Q, respectively.

The digital I and Q signals are applied to a first Doppler filter bank 24, which is a conventional Doppler filter bank for filtering the return signals into various frequency bins. It should be noted that the term “Doppler filter bank” is meant to include all manifestations of such processing, whether accomplished by dedicated hardware or by computation in a processor.

The digital I and Q signals are also applied to a low-velocity target enhancer or target detection enhancer 30. Enhancer 30 includes the cascade of a wavelet filter 32I, a data compressor 34I, and an inverse wavelet filter 36I for the I-channel digital signal returns, and a wavelet filter 32Q, a data compressor 34Q, and an inverse wavelet filter 36Q. Enhanced target I and Q information is produced on output signal paths 381 and 38Q, respectively, and is applied to a second Doppler filter bank 40. Doppler filter bank 40 routes the enhanced signals into frequency bins, as known in the art. A wavelet processor differs from a Doppler processor in that the dimensions are different. The Doppler processor channelizes by time and frequency range, while the wavelet processor channelizes by translation and scale.

According to an aspect of the invention, there is limited or no overlap between the Doppler frequencies of the bins of Doppler filter banks 24 and 40. That is to say that the highest-frequency bins of Doppler filter bank 40 are at frequencies lower, or at least no greater, than the lowest-frequency bins of Doppler filter bank 24.

In operation of system 10 of FIG. 1, the low-velocity target enhancer 30 uses wavelet filtering to translate the returns into different dimensions than frequency and time. More particularly, the dimensions of the wavelet filtering are translation and scale. The properties of wavelet filters are well known, and are described, for example, at http://users.rowan.edu/%7Epolikar/WAVELETS/WTtutorial.html Any filter decomposes the content of input signals into orthonormal components. Conventional Doppler filters decompose signals into their sinusoidal frequency components with the presumption that target information can be extracted from a combined target and interference signal mix by virtue of differences in frequency content. Wavelet processing decomposes signals into scale (not frequency) of wavelet functions, where the wavelet serves as the analog of a sine wave at a particular frequency in Doppler filtering. All wavelet filters apply specific mother wavelet functions tailored to produce beneficial signal decomposition for the particular signal processing problem at hand. The scaling of the mother wavelet function, the contraction or dilation in time of the fundamental wavelet function, is akin to frequency translation in a Doppler filter bank. Whereas Doppler filtering is constrained to fundamental sine wave functions, wavelet filtering can apply a variety of mother wavelet functions to decompose time varying signals.

For radar signals composed of target and interference sources such as clutter, wavelet filtering produces inherently finer separation of signal components than Doppler filters in the lower portion of the Doppler spectrum. The lower portion of the Doppler spectrum is comprised of low velocity targets and clutter. This finer separation or resolution property is exploited in this invention to arrive at enhanced low velocity target extraction from clutter.

The wavelet filtering inherently provides better frequency resolution in the low-frequency domain than a Doppler filter. On the other hand, the resolution at higher frequencies is less than that of a Doppler filter. According to an aspect of the invention, wavelet filtering is performed on the I and Q components, and the resulting wavelet filtered information is data compressed in blocks 34 i and 34Q by editing or setting to zero the first M filter outputs, which is to say the first M of N signals to exit the filter, where N represents the total number of processed radar pulses in a pulse sequence. The value of M can be tailored to suit the properties of the clutter. The data compressed I and Q components from blocks 34I and 34Q are applied to inverse wavelet filters 36I and 36Q, which reconvert the data compressed returns to the frequency-time domain.

The output signals from Doppler filters 24 and 40 are at different frequencies. The two banks of Doppler filters may be viewed as being two portions of a single bank of Doppler filters covering the combined frequency range. However, the wavelet processing which precedes the filtering by Doppler filter bank 40 tends to suppress clutter to provide better target resolution than would be obtainable by use of Doppler filters 24 and 40 without the wavelet processing.

The various frequency bins of the Doppler filters 24 and 40 of FIG. 1 may be viewed as being the bins of a single, wideband Doppler filter banks, so the outputs of Doppler filter banks 24 and 40 may be combined, and used for further processing in conventional manner.

Thus, a method according to an aspect of the invention is for filtering coherent multipulse radar return signals. The method comprises the steps of separating (20) the return signals into mutually orthogonal I and Q components, performing wavelet processing (32I) on the I component to generate an I wavelet filter output, and performing wavelet processing (32Q) on the Q component to generate a Q wavelet filter output. The I and Q wavelet filter outputs are data compressed (34 i, 34Q) to remove those portions which are most likely to contain clutter, to thereby generate data compressed wavelet filter outputs. An inverse wavelet filtering function (36I, 36Q) is performed on the compressed wavelet filter outputs, to thereby generate edited coherent I and Q components. Doppler filtering (40) (DFT or FFT) is performed on the edited coherent I and Q components to thereby generate frequency-domain target data in which the minimum discernible target velocity is improved.

In a particular mode of the method according to this aspect of the invention, the step of data compressing (34I, 34Q) the I and Q wavelet filter outputs comprises the step of removing those portions of the I and Q wavelet filter outputs which have the largest amplitudes.

In another particular mode of the method according to this aspect of the invention, the step of data compressing (34I, 34Q) the I and Q wavelet filter outputs comprises the step of removing those I and Q wavelet filter outputs which occur earliest in the stream of wavelet responses.

In yet another particular mode of the method, the additional step is performed of Doppler filtering (24) the mutually orthogonal I and Q components into frequency bins having frequencies above the frequencies associated with the minimum discernible target velocity.

Patent Citations
Cited PatentFiling datePublication dateApplicantTitle
US5602760 *Feb 2, 1994Feb 11, 1997Hughes ElectronicsImage-based detection and tracking system and processing method employing clutter measurements and signal-to-clutter ratios
US6097669 *Jul 1, 1999Aug 1, 2000Jordan; James R.Wavelet filtering of sodar signals
Non-Patent Citations
Reference
1 *Colestock, M. A., "Wavelets- a new tool for signal processing analysts", Proceedings of 1993 IEEE.AIAA 12th Digital Avionics Systems Conference, Forth Worth, TX USA Oct. 25-28, 1993.
Referenced by
Citing PatentFiling datePublication dateApplicantTitle
US8447472 *Jan 16, 2007May 21, 2013Ford Global Technologies, LlcMethod and system for impact time and velocity prediction
US8567251Sep 22, 2008Oct 29, 2013Vega Grieshaber KgDetailfunction based measurement
US9188669 *May 10, 2013Nov 17, 2015Furuno Electric Company LimitedTarget finder, high resolution processing device, and high resolution processing method
US20080172156 *Jan 16, 2007Jul 17, 2008Ford Global Technologies, Inc.Method and system for impact time and velocity prediction
US20100307251 *Sep 22, 2008Dec 9, 2010Roland WelleDetailfunction Based Measurement
US20130300597 *May 10, 2013Nov 14, 2013Furuno Electric Company LimitedTarget finder, high resolution processing device, and high resolution processing method
US20150185244 *Dec 23, 2014Jul 2, 2015Kabushiki Kaisha Toyota Chuo KenkyushoRadar apparatus and method of determining sign of velocity
CN101835045A *May 5, 2010Sep 15, 2010哈尔滨工业大学Hi-fidelity remote sensing image compression and resolution ratio enhancement joint treatment method
CN101861525BSep 22, 2008Aug 28, 2013Vega格里沙贝两合公司Detailfunction based measurement
WO2009037000A3 *Sep 22, 2008May 14, 2009Grieshaber Vega KgDetailfunction based measurement
Classifications
U.S. Classification342/160, 342/195, 342/162, 342/159
International ClassificationG01S13/52
Cooperative ClassificationG01S2007/2886, G01S7/2923, G01S13/522
European ClassificationG01S7/292C, G01S13/522
Legal Events
DateCodeEventDescription
May 26, 2005ASAssignment
Owner name: LOCKHEED MARTIN CORPORATION, MARYLAND
Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNOR:FERRANTE, JOHN G.;REEL/FRAME:016619/0859
Effective date: 20050525
Nov 2, 2009FPAYFee payment
Year of fee payment: 4
Dec 13, 2013REMIMaintenance fee reminder mailed
May 2, 2014LAPSLapse for failure to pay maintenance fees
Jun 24, 2014FPExpired due to failure to pay maintenance fee
Effective date: 20140502